Acupuncture is often considered for the treatment of chronic pain, but its value remains controversial. Many good quality meta-analyses have been published synthesizing randomized trials of acupuncture for back pain, headache and osteoarthritis. However, these have typically come to rather indeterminate conclusions, largely because of the questionable size and quality of the included trials: a typical trial has perhaps 30 patients per group and did not specify a method of allocation concealment. The landscape of clinical research in acupuncture has recently been dramatically altered by the completion of several large, high quality trials. For example, the PI of the current proposal has published a randomized trial with 401 patients;the GERAC group has completed three trials with close to 1000 patients each, and the four ARC trials have randomized close to 10,000 patients in total. All of these trials are of high methodologic quality. The incorporation of these recent large trials of acupuncture into meta-analysis is therefore a high priority of acupuncture research. We propose obtaining individual patient level data from randomized trials of acupuncture for chronic pain. These data will be combined to create an individual patient level data set. This data set will be analyzed to address a variety of objectives. One of our main interests concerns the effectiveness of acupuncture. We will analyze the data, separately for each type of chronic pain condition, to determine whether acupuncture is more effective than sham acupuncture and whether it is superior to usual care control. Our data set will also allow us to answer a variety of other questions including: do different types of sham control differ? What characteristics of acupuncture, such as treatment style or number of treatment sessions, are associated with good clinical outcomes? Do the effects of acupuncture differ by type of pain? Our results have important implications both for the clinical management of chronic pain and for research design in acupuncture. After publication of the principal results of these analyses, a de-identified raw data set will be posted on an Open Access website for the benefit of the acupuncture research community as a whole. The grant proposal is in the form of a collaboration - the Acupuncture Trialists'Collaboration - between the PI's of major acupuncture clinical trials, all of whom have formally agreed to take part. PUBLIC HEALTH RELEVANCE: Both the assessment of the clinical value of acupuncture, and research design of acupuncture studies, depend critically on the careful synthesis of prior research data. We plan to combine raw data from all large, high-quality randomized trials of acupuncture for chronic pain into a single data set to enable individual patient data meta- analysis. Most of these studies were published recently and have not been incorporated into published meta-analyses.